Mini-Project 02
Data Visualization Project 02
Importing Data Set
Here I am importing the spacex.csv data set into the markdown file, and initializing my libraries. This data set looks at launch data taken from 2010-0218. It contains information including date, launch site, payload size, mission outcome and more.
library(tidyverse)
library(sf)
library(plotly)
library(dplyr)
spacex <- read_csv("../data/spacex.csv") %>%
mutate(
Date = as.Date(Date, "%B %d, %Y"), # e.g. "June 4, 2010"
Year = year(Date), # extract integer year
success = (`Mission Outcome` == "Success"), # TRUE/FALSE
payload_kg = parse_number(`Payload Mass (kg)`) # strip commas, make numeric
)
glimpse(spacex)## Rows: 51
## Columns: 14
## $ `Flight Number` <chr> "1", "2", "3", "4", "5", "6", "7", "8", "9", "10",…
## $ Date <date> 2010-06-04, 2010-12-08, 2012-05-22, 2012-10-08, 2…
## $ `Time (UTC)` <time> 18:45:00, 15:43:00, 07:44:00, 00:35:00, 15:10:00,…
## $ `Booster Version` <chr> "F9 v1.0", "F9 v1.0", "F9 v1.0", "F9 v1.0", "F9 v1…
## $ `Launch Site` <chr> "CCAFS LC-40", "CCAFS LC-40", "CCAFS LC-40", "CCAF…
## $ Payload <chr> "Dragon Spacecraft Qualification Unit", "Dragon de…
## $ `Payload Mass (kg)` <chr> NA, NA, "525 ", "500 ", "677 ", "500 ", "3,170 ", …
## $ Orbit <chr> "LEO", "LEO", "LEO", "LEO", "LEO", "Polar orbit", …
## $ Customer <chr> "SpaceX", "NASA (COTS) NRO", "NASA (COTS)", "NASA …
## $ `Mission Outcome` <chr> "Success", "Success", "Success", "Success", "Succe…
## $ `Landing Outcome` <chr> "Failure (parachutes)", "Failure (parachutes)", "N…
## $ Year <dbl> 2010, 2010, 2012, 2012, 2013, 2013, 2013, 2014, 20…
## $ success <lgl> TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TR…
## $ payload_kg <dbl> NA, NA, 525, 500, 677, 500, 3170, 3325, 2296, 1316…Data Summary
In this section I filtered the SpaceX data to see: - Number of launches per year - Success rate per launch site - Average payload mass by booster version
#Launches per year
launches <- spacex %>%
group_by(
Year = as.integer(format(as.Date(Date, "%B %d, %Y"),"%Y"))
) %>%
summarize(
`Number of Launches` = n(),
`Success Rate` = mean(success, na.rm = TRUE)
) %>%
arrange(Year)
launches## # A tibble: 8 × 3
## Year `Number of Launches` `Success Rate`
## <int> <int> <dbl>
## 1 2010 2 1
## 2 2012 2 1
## 3 2013 3 1
## 4 2014 6 1
## 5 2015 7 0.857
## 6 2016 8 1
## 7 2017 18 1
## 8 2018 5 0.8#Success rate
success <- spacex %>%
group_by(`Launch Site`) %>%
summarize(
`Total Launches` = n(),
`Success Rate` = mean(`Mission Outcome` == "Success", na.rm = TRUE),
.groups = "drop"
) %>%
arrange(desc(`Total Launches`))
success## # A tibble: 5 × 3
## `Launch Site` `Total Launches` `Success Rate`
## <chr> <int> <dbl>
## 1 CCAFS LC-40 26 0.962
## 2 KSC LC-39A 13 1
## 3 VAFB SLC-4E 8 1
## 4 CCAFS SLC-40 3 1
## 5 CCAFS SLC-40 (after static fire on KSC LC-39A) 1 0#Avg Mass
payload <- spacex %>%
group_by(Booster = `Booster Version`) %>%
summarise(
`Average Payload (kg)` = mean(parse_number(`Payload Mass (kg)`), na.rm = TRUE),
`Number of Launches` = n(),
.groups = "drop"
) %>%
arrange(desc(`Average Payload (kg)`)) %>%
filter(!is.na(`Average Payload (kg)`))
payload ## # A tibble: 31 × 3
## Booster `Average Payload (kg)` `Number of Launches`
## <chr> <dbl> <int>
## 1 F9 B4 B1041.1 9600 1
## 2 F9 FT B1029.1 9600 1
## 3 F9 FT B1036.1 9600 1
## 4 F9 FT B1036.2 9600 1
## 5 F9 FT B1037 6761 1
## 6 F9 B4 B1044 6092 1
## 7 F9 FT B1034 6070 1
## 8 F9 FT B1030 5600 1
## 9 F9 FT B1021.2 5300 1
## 10 F9 FT B1020 5271 1
## # ℹ 21 more rowsInteractive Data Visualization
spacex <- spacex %>%
mutate(payload_kg = parse_number(`Payload Mass (kg)`))
plot_ly(
data = spacex,
x = ~as.numeric(`Flight Number`),
y = ~payload_kg,
size = ~payload_kg,
color = ~`Launch Site`,
text = ~paste0(
"Flight #: ", `Flight Number`, "<br>",
"Date: ", Date, "<br>",
"Booster: ", `Booster Version`, "<br>",
"Site: ", `Launch Site`, "<br>",
"Payload: ", payload_kg, " kg<br>",
"Outcome: ", `Mission Outcome`
),
hoverinfo = "text",
mode = "markers"
) %>%
layout(
title = "SpaceX Payload by Flight",
xaxis = list(title = "Flight Number"),
yaxis = list(title = "Payload Mass (kg)"),
legend = list(title = list(text = "<b>Launch Site</b>"))
)Spatial Visualization
world_shapes <- read_sf("../data/ne_110m_admin_0_countries/ne_110m_admin_0_countries.shp", quiet = TRUE) %>%
filter(ISO_A3 == "USA")
#site coords
site_coords <- tribble(
~site, ~lon, ~lat,
"CCAFS LC-40", -80.577366, 28.561871,
"CCAFS SLC-40", -80.583333, 28.583330,
"KSC LC-39A", -80.604333, 28.608389,
"VAFB SLC-4E", -120.610829, 34.632092
)
success_by_site <- spacex %>%
group_by(site = `Launch Site`) %>%
summarise(
total_launches = n(),
success_rate = mean(`Mission Outcome` == "Success", na.rm = TRUE),
.groups = "drop"
)
sites_sf <- success_by_site %>%
left_join(site_coords, by = "site") %>%
filter(!is.na(lon), !is.na(lat)) %>%
mutate(
lon = lon + runif(n(), -0.1, 0.1),
lat = lat + runif(n(), -0.1, 0.1)
) %>%
st_as_sf(coords = c("lon", "lat"), crs = 4326)
#Plot
ggplot() +
geom_sf(data = world_shapes, fill = "gray95", color = "gray80") +
geom_sf(
data = sites_sf,
aes(size = total_launches, color = success_rate),
alpha = 0.8
) +
scale_size(range = c(1, 5)) +
scale_color_viridis_c(labels = scales::percent_format(accuracy = 1)) +
labs(
title = "SpaceX Launch Sites in the USA (2010–2021)",
subtitle = "Point size = total launches, color = success rate",
size = "Total\nLaunches",
color = "Success\nRate"
) +
theme_minimal(base_size = 14) +
theme(
panel.grid.major = element_line(color = "white"),
plot.title = element_text(face = "bold", size = 16),
plot.subtitle = element_text(size = 12)
)
Visualization of a Model
library(broom) # for tidy()
library(dotwhisker) # for dwplot()
# avoid lubridate::year() name clash
spacex2 <- spacex %>%
rename(launch_year = Year)
# fit the logistic model
log_mod <- glm(
success ~ payload_kg + launch_year,
data = spacex2,
family = binomial
)
ggplot(spacex, aes(x = payload_kg, y = as.numeric(success))) +
geom_jitter(
aes(x = payload_kg,
y = as.numeric(success),
colour = success), # <-- map colour to the success flag
height = 0.02,
alpha = 0.6,
size = 2
) +
scale_colour_manual(
"Launch Outcome",
values = c(
"TRUE" = "forestgreen",
"FALSE" = "firebrick"
)
) +
stat_smooth(
method = "glm",
method.args = list(family = "binomial"),
se = TRUE,
color = "forestgreen",
fill = alpha("palegreen", 0.3)
) +
scale_y_continuous(labels = scales::percent_format()) +
labs(title = "Success Probability vs. Payload Mass",
x = "Payload Mass (kg)",
y = "Probability of Success") +
theme_minimal()
ggsave("figures/success_prob_vmass.png",
width = 6, height = 4, units = "in", dpi = 300)
tidy(log_mod, conf.int = TRUE) %>%
filter(term != "(Intercept)") %>%
dwplot(point_size = 3) +
labs(title = "Predictors of Launch Success",
x = "Log-Odds Coefficient",
y = "") +
theme_minimal()
The first visualization (“Success Probability vs. Payload Mass”) shows a fitted logistic curve overlaid on the raw jittered success/failure points. Notice how the blue line sits up near 100 % success for almost the entire range of payloads and only very slightly dips at the highest masses. In other words, payload mass really doesn’t effect whether a flight succeeds or not (the 95 % confidence ribbon is very wide at the far end but the mean curve is essentially flat).
The second visualization (“Predictors of Launch Success”) is a classic coefficient‐plot of your glm. It shows the log-odds estimates (with confidence intervals) for both payload_kg and launch_year, excluding the intercept. Both coefficients sit close to zero and their intervals overlap zero, indicating neither payload nor year is a strong predictor of success.